Models & assistants Model or AI service
BharatGen publishes Param-1 for Hindi and English text
Param-1 is a 2.9B language model trained from scratch on Hindi and English. BharatGen publishes an early instruction-tuned checkpoint for further study. [1][2]
- Model access
- Downloadable Param-1 2.9B Instruct early checkpoint. [2]
- License
- Apache License 2.0 for the cited checkpoint. [2]
What the model does
BharatGen’s July 2025 report describes the bilingual training and early evaluations. The checkpoint documentation identifies it as an early supervised fine-tuning release. [1][2]
Why it matters
A bilingual model at this scale gives researchers a way to study how tokenizer design and training data affect Hindi and English tasks. [1][2]
Who can use it
Researchers can evaluate the public checkpoint for bilingual question answering and related text applications. The card describes it as a starting point for more work. [2]
Key dates
- Param-1 report first submitted [1]
Ways to use it
Test Hindi, English and mixed-language prompts with verifiable expected answers. Compare accuracy across those cases and record the exact checkpoint used. [1]
Model limitations
The published model is an early text checkpoint. Its Hindi and English training does not establish support for every language in India or readiness for high-impact workflows. [1][2]
Sources and verification
Catalog evidence checked How we verify records
- Primary source BharatGen’s first Param-1 2.9B report and submission history
- Primary source bharatgenai/Param-1-2.9B-Instruct
- Other source PARAM-1 BharatGen 2.9B Model
Related profiles
Models & assistants
GPT-6.2 Astra and Sol
OpenAI
Astra targets demanding projects. Sol targets everyday tasks. Both support text and image input in the API.
Models & assistants
MAI Code 2 and Code 2 Flash
Microsoft AI
Code 2 targets complex software work. Flash targets frequent edits, explanations and routine fixes.
Models & assistants
MAI Realtime 1
Microsoft AI
Users can try a MAI Labs demo. Developers can register for a limited Foundry preview.